PhiC31 integration efficiency

In the selection-based integration assays the efficiency of PhiC31 integrase to integrate plasmid DNA into the mammalian genome was determined relative to the quantity of G418- resistant colonies. The integration efficiency is predominantly influenced by the chromosomal context due to the intrinsic influence of the genomic DNA flanking the integration site, on the promoter and the transgene within the substrate plasmid. The choice and design of the substrate plasmid and in particular the promoter influences efficiency and transgene expression. The huge diversity of the chromosomal context within the genome derived from different cell lines and organs reasoned the inconsistent integration efficiency among different cell lines (Section 4.5).This is supported by the finding that chromatin conformation and

accessibility were suggested to influence PhiC31 integrase pseudo site selection (Portlock and

Calos, 2003) and transgene expression levels. The frequency of integration by a related HIV integrase was shown to be influenced not only by the structure of the target site, but also by DNA curvature and flexibility (Pruss et al., 1994), thus proposing that features may also affect the PhiC31 integrase target site selection. Another parameter which influences integration efficiency refers to the transfection method (delivery into the nucleus). Since equal transfection conditions via lipofection were assumed throughout the experiments, beneficial effects on plasmid delivery, e.g. addition of nuclear localisation signals (NLS), are not considered further within the context of this work. The overall integration efficiency upon

lipofection-mediated transfection in vitro is generally low, but still higher than integration by

homologous recombination (10-6) for most mammalian cells (Vega, 1991). PhiC31 integrase-

mediated integration naturally accomplished an approximate 100-fold improved

recombination rate in vitro compared to random integration (Chalberg et al., 2006).

The constitution and the proficiency of the particular integrase likely play the most pivotal role in respect to integration efficiency. Since the crystal structure of the PhiC31 integrase has not been solved yet, homology modelling and secondary structure prediction (Yang and Steitz, 1995; Li et al., 2005) could support site-directed mutational analysis addressing recombination activity and in particular integration efficiency.

In the initial screen (Figure 4.7), five mutants showed a slight improvement in integration efficiency between 1.2- and 1.7-fold, assuming a mild functional preference for alanine

instead of the large side chains of the charged amino residues (detailed structure in appendix). Single amino acid changes at these critical residues take certain effects on the overall enzymatic functionality e.g. by altering the charge and the conformation. However, replacing large side chains of amino acids, the methyl group of the small amino acid alanine does not necessarily lead to increased recombination efficiency.

Seven mutants showed a fivefold decline of integration efficiency suggesting structural or functional importance associated with the native charged amino acid positions (Figure 4.7). The small side group of alanine abolishes integration activity likely at these particular positions.

Ten mutants showed rather similar integration efficiency compared to wt integrase. Therefore, these amino acids were not critical to maintain the recombination activity at the level of wt integrase. Although the goal of this study targets to increase integration efficiency, it would be of great interest to determine, at which recombination step the mutant proteins are deficient, to gain more information and understanding about structure-function relationships. Initial experiments addressing synapsis, DNA cleavage and control of directionality during the PhiC31 integrase-mediated recombination reaction might gain additional knowledge (Thorpe et al., 2000; Smith et al., 2004).

According to the secondary structure prediction (Figure 1.5) the DNA binding domain comprises three repetitive β-sheets within position D366-R392 and a long α-helix comprising position between M413 and E480. A putative coiled-coil region, which is a common structural motif in proteins, in length of about 28 amino acids with a periodicity of four

heptads forming several supercoiled α-helices has also been observed within the binding

domain (Rowley et al., 2008; McEwan et al., 2009). In the first β-sheet comprising amino acid sequence 365-MDKLYC-370, two mutants (D366A and K367A) were created in this study, showing totally diverse effects on integration efficiency. Removal of the acidic carboxy group (COOH) of aspartic acid (D) and replacing it with the methyl group gained slightly integration efficiency. The mutant K367A dropped integration activity about fivefold when the positive charge is removed. This might be due to the substitution of a large basic amino group by the small methyl group, which likely implies the importance of the polar side chain at that position to maintain hydrogen bonds between individual side chains of amino acids. The two mutants (E382A and E383A) showed both an approximate 1.3-fold increase. The

residues are located between two connecting β-sheets. Mutations towards alanine at these

positions might improve orientation and binding affinities between the second and the third β- sheet. The side-chain of the positively charged amino acid arginine (R) at position 380 might

be involved to maintain the β-sheet structure of the second sheet, since alanine reduced activity dramatically. The third beta-sheet included two amino acids changed to alanine; K386A and R390A. The mutation effects resulted in a drop to 60 % integration efficiency compared to the wt integrase. Removal of positively charged side chains might consequently abolish the disruption of native interaction between the sheets, as indicated by the fivefold loss of recombination activity in the mutants R393A and R394A. These findings suggest that positively charged amino acids found in structures forming β-sheets support stronger binding to negatively charged DNA.

The most beneficial mutant D470A lies within the structural motif of a coiled-coil region

between two alanine residues (Figure 4.5), in which four α-helices are coiled together in

heptamers (Rowley et al., 2008; McEwan et al., 2009). By replacing the negatively charged side chain of aspartic acid (D) with alanine the protein interaction interface likely performs

enhanced substrate recognition or improved α-helical propensity. The putative coiled-coil

motif is strongly involved in substrate recognition and consequently in attB_×attP synapsis

(McEwan et al., 2009). The approximate tenfold decrease of mutants D417A and E432A suggests that the negatively charged sidechains of aspartic acid (D) and glutamic acid (E) are likely involved in protein active sites or cationic bonds rather than in DNA interaction. Whether or not the mutated residues in the double mutants are in close proximity in its active conformation could not be determined. Mutants showing improved efficiency were likely altered in specificity as well. However, it was not investigated whether and when integration efficiency is associated with integration specificity, assuming that no obvious coherency exists.